Abstract
Refractory State Coding (RSC) is a new coding strategy based on a functional model of the stimulated neural population. Our hypothesis is that RSC stimulation would convey the information contained in acoustic signals more effectively, improving sound perception and hearing performance for speech in noise and music. Two main factors that RSC takes into account are channel interaction [1] and refractory properties [2] of the stimulated neural population. They can be characterized by electrophysiological measurements of the evoked compound action potential (ECAP) using “spread of excitation” (SoE) and “recovery function” characterization paradigms respectively [3]. Using this information, for a given stimulus sequence, it is possible to calculate the refractory state of each stimulation site at any given time. In RSC, the stimulus is shaped according to the refractory states of stimulation sites. The spectral representation of the input sound is weighted by the refractory recovery information as well as the electric field distribution function before the next stimulus is selected. The Nucleus 24 and Nucleus Freedom family of cochlear implants incorporate Neural Response Telemetry (NRT) circuitry which is able to conveniently measure the ECAP from the implanted intracochlear electrodes, allowing the model to be custom-fitted to a patient. A software implementation of the standard ACE strategy for the Nucleus Cochlear Implant system is available in the Nucleus Matlab Toolbox. We implemented the RSC strategy in a compatible fashion in Matlab.
Babacan, O